Employment of spot sprayers will lead to less herbicide consumption and less hazard. Because of high importance attached to weed control in agriculture and to its related expenses, plenty of research is being carried out in this field. In this study 300 digital images were taken and developed from different agricultural fields in natural outdoor conditions. Images possessed a resolution of 1200×1600 pixels in reference to a field of view of 70×60 cm2. Constitutional resolution as related to the composition of three main color components in plants as well as in 7 common weed species (in Fars sugar beet fields) were extracted using discriminate analysis. Based upon the obtained relations, a suitable algorithm has been proposed through which, one is able to discriminate weeds from sugar beet plants in any form of light condition regardless of whether they are in bright light or in the shadow of either other leaves or vehicle. Correct classification rate for each one of the seven weed plants has been evaluated and discussed. Minimum correct classification rate was related to Convolvulus arvensis L. in shadow (70.8%) while maximum related to Setaria veridis L. Beauv (95.2%). Correct classification rate in the final segmentation algorithm was found to be88.5% in sunlight and 81.8% in shadow.
)